Abstract— Since displays are planar and with a limited field of view, to visualize 360º (spherical) content, it is necessary to employ a projection to map pixels on the sphere to a 2D plane segment. This 2D plane is called viewport and is created with some limited field of view, usually much less than 360º. To create the viewport, 3D points on the sphere are projected to the 2D plane usually with a perspective projection. This process leads to geometric distortions in the viewport, such as objects that appear stretched or image structures that are bent. This paper proposes a content-dependent objective quality assessment procedure to evaluate line distortions that occur during the viewport creation process, to identify which projection center minimizes the subjective impact of these distortions. To achieve this objective, features that characterize the amount of line distortion in the viewport image are extracted and used by a Support Vector Machine (SVM) classifier, to obtain the viewport quality. To train the classifier, a subjective evaluation of rendered viewport images was conducted to obtain the perceptual scores for different types of content and projection centers. The experimental results show that the proposed metric is able to predict the viewport quality with an average accuracy of 91.2%.